# Sesame (*Sesamum indicum*) — Literature Survey

- **ResearchForge workspace:** `research/sesame-seeds/`
- **Date:** 2026-08-26
- **Depth:** comprehensive — 28 query variants, 4 scholarly sources, 5 citation expansions
- **Status:** metadata-level survey only. Screening decisions, full-text acquisition, extraction acceptance, analysis-method selection, and final claims remain human-gated.

## 1. Method and corpus

- Queried OpenAlex, Crossref, arXiv, and Semantic Scholar for biology/genetics, germination, planting, photoperiod, climate, water, soil, fertility, pests/diseases/weeds, capsule development, dehiscence/shattering, harvest, threshing/cleaning, storage, oil extraction, yield/economics, Mexico, and food-safety boundaries.
- The final canonical batch contains **2,030 library records** and **1,585 deduplicated records**, including **1,268 unique DOIs**. Of the deduplicated records, 317 lack a DOI, 76 lack a year, and 779 lack an abstract.
- Accumulated raw-source coverage reports 700 arXiv, 699 Crossref, 660 OpenAlex, and 358 Semantic Scholar records; raw-file DOI coverage (1,529) is broader than the final canonical deduplicated export because this workspace preserves multiple batch attempts.
- A corrected `search resume` recovered 82 records but the installed CLI wrote only those records to `results.jsonl`, leaving the deduplicated file stale. Those records are preserved separately as `resume-results.jsonl`; they were **not silently merged** into the canonical corpus. The original 28-query batch was rerun to restore a self-consistent canonical `results.jsonl`/`results-deduped.jsonl` pair.
- Generated `queue.csv` with **1,998 pending records** and zero decisions. No screening decision was made.
- The analysis requires sesame identity plus crop-domain context and filters obvious SESAME software acronyms, other-crop applications of sesame products, wrong *Sesamum* species, and biomedical noise from agronomic topic counts.
- Titles, abstracts when present, identifiers, and citation metadata establish candidate existence and coverage only. No numerical agronomic recommendation is asserted.

## 2. Bottom line

The strongest metadata threads are **capsule dehiscence/shattering**, cultivar/genetic diversity, germination and establishment, drought/waterlogging/salinity responses, nitrogen management, stored-seed quality, and seed-oil processing. Sesame must be treated as two linked but separate outputs: **clean seed** and **extracted oil**.

The weakest human-scale evidence concerns cutting/bundling, hand threshing, winnowing, cleaning losses, labor-hours per kilogram, and small-press settling/filtration economics. Capsule-development/maturity evidence is also much thinner than shattering genetics. Mexico evidence is limited but not absent: one direct genotype × environment study in Guerrero is present.

## 3. Candidate evidence by topic

### 3.1 Identity, diversity, and breeding

- Global transcriptome and EST-SSR markers — `10.1186/1471-2164-12-451` (2011).
- Deep resequencing and allelic variation — `10.1186/s12870-014-0225-3` (2014).
- Morphological/genetic diversity among accessions — `10.1007/s12298-015-0322-2` (2015).
- “Omics” review candidate — `10.3389/fpls.2017.01154` (2017).

### 3.2 Germination, emergence, planting, and flowering

- Planting-date effects on growth/development — `10.2134/agronj1987.00021962007900040024x` (1987).
- Temperature and presoaking effects on germination — `10.1016/0098-8472(95)00013-x` (1995).
- Hydrothermal-time germination response to temperature/water potential — `10.1007/s11738-017-2549-8` (2017).
- Temperature, soil bulk density, and sowing-depth emergence model — `10.1016/j.indcrop.2019.111770` (2019).
- Row-spacing response in semi-arid regions — `10.13189/ujar.2020.080402` (2020).
- Flowering response to temperature and photoperiod — `10.1071/ar9921101` (1992). Citation expansion succeeded with 31 edges.
- Photoperiod-response gene candidate — `10.1186/s12870-018-1583-z` (2018).

### 3.3 Climate, water, soil, and salinity

- Waterlogging response patterns in contrasting genotypes — `10.1371/journal.pone.0149912` (2016).
- Drought-response comparison of contrasting genotypes — `10.1038/s41598-017-09397-6` (2017).
- Contrasting drought/waterlogging methylation programs — `10.1016/j.plantsci.2018.09.012` (2018).
- Salt-stress pathways in contrasting genotypes — `10.1186/s12870-019-1665-6` (2019).
- Genotype response across irrigation levels — `10.1002/agj2.20167` (2020). Citation expansion succeeded with 44 edges.
- Drought-stress irrigation productivity candidate — `10.1016/j.agwat.2022.107569` (2022).

These are stress/experimental candidates, not practical household irrigation schedules. Seasonal totals remain `Unknown`.

### 3.4 Fertility

- Nitrogen rates, efficiency, and cultivar seed-yield response — `10.1080/00103624.2018.1510949` (2018).
- Nitrogen fertilization and growth/seed yield — `10.1155/2019/5027254` (2019).
- Sulphur and nutrient uptake/seed yield — `10.1007/s12517-021-07229-6` (2021).
- Biochar addition, sesame growth/yield, mineral nutrients, and soil properties — `10.3390/agronomy9020055` (2019).

### 3.5 Pests, diseases, and weeds

- Sesame phyllody phytoplasma — `10.1590/s1982-56762010000500006` (2010).
- Sesame seed bug and harvest loss — `10.4314/sinet.v31i2.66645` (2011).
- Charcoal-rot resistance QTL — `10.1038/s41598-017-08858-2` (2017).
- Sowing date/insecticide frequency for sesame webworm — `10.1186/s40066-018-0190-4` (2018).
- Weed-management extension candidate — `10.32473/edis-ag396-2021` (2021).

Local Mexico pest calendars and household integrated-management guidance remain gaps.

### 3.6 Capsule development, dehiscence, and shattering

- Capsule dehiscence anatomy — `10.1017/s0021859699007364` (2000).
- Inheritance of indehiscent capsule character — `10.15832/ankutbd.539005` (2019).
- Growth-regulator treatments and capsule dehiscence — `10.1016/s0378-4290(99)00076-3` (2000).
- High-density SNP map and capsule-shattering trait mapping — `10.3389/fpls.2021.679659` (2021).
- Capsule/seed development candidate — `10.9734/ijecc/2020/v10i430189` (2020).
- Review of breeding history for shattering — `10.1007/s11033-022-07636-2` (2022). This earlier-run citation-expansion anchor produced 50 edges but is not present in the final canonical deduplicated export; topic pages must not use it as a canonical candidate.
- Dehiscent vs indehiscent capsule metabolite/mechanism candidate — `10.1016/j.fochms.2024.100231` (2024).

Capsule type, maturity stage, harvest timing, and measured seed loss must be recorded together. Evidence from non-dehiscent mechanized cultivars cannot be transferred to dehiscent household seed without cultivar context.

### 3.7 Harvest, threshing, and cleaning

- Sowing date × harvest time candidate — `10.20546/ijcmas.2023.1202.011` (2023).
- Harvest-aid drydown/maturity candidate — `10.5772/intechopen.91011` (2020).
- Sesame-thresher development/evaluation — `10.31018/jans.v13isi.2824` (2021).
- Adaptation of threshing equipment — `10.1088/1755-1315/1290/1/012015` (2024).
- Adaptation of grain-cleaning equipment — `10.3390/agriengineering7060179` (2025).

These machinery candidates do not answer household cutting, bundling, hand threshing, winnowing, labor, and loss questions. Those remain `Unknown`/Needs experiment.

### 3.8 Drying, storage, viability, and stored pests

- Storage life under humid tropical conditions — `10.15258/sst.2008.36.2.11` (2008).
- Physiological seed quality during storage — `10.1590/s1806-66902014000100017` (2014).
- Environmental factors, storage periods, quality, and longevity — `10.1186/s43170-024-00247-w` (2024). Citation expansion succeeded with 37 edges.
- Storage technologies, pest infestation, and postharvest loss — `10.1016/j.jspr.2023.102161` (2023).
- Stored-seed insect dynamics/losses — `10.1007/s42690-024-01179-8` (2024).

Planting-seed viability, edible-seed quality, insect loss, mold/mycotoxin risk, and oil oxidation are different endpoints and must not be merged.

### 3.9 Seed oil and processing

- Roasting effects on oil composition/quality — `10.1002/jsfa.2740500413` (1990).
- Expeller wormshaft speed/moisture and oil/cake quality — `10.1002/ts.115` (2003).
- Low-temperature screw-press extraction optimization — `10.1007/s11947-017-1885-4` (2017). Citation expansion succeeded with 50 edges.
- Cold pressing and oil/cake quality — `10.3390/foods10092108` (2021).
- Processing/storage and sesame-oil stability — `10.1007/s11746-997-0159-0` (1997).

Manual/small-press throughput, filtered-oil recovery, settling/filtration, residual cake handling, and labor remain `Unknown`.

### 3.10 Yield and economics

- Tillage, sesame seed yield, energy budget, and economic return — `10.3906/tar-1011-1505` (2012).
- Production challenges/opportunities review for Myanmar — `10.3390/su12093515` (2020).
- Uganda production/export trends — `10.1051/ocl/2020073` (2021).
- Yield/quality research review — `10.3389/fpls.2022.863521` (2022).
- Smallholder technical efficiency — `10.1007/s43546-023-00604-3` (2023).

All field-yield values require cultivar/capsule type, environment, density, water/fertility, maturity, harvest method, shattering loss, and clean-seed definition. Seed yield and extracted-oil yield remain separate.

## 4. Mexico and Latin America

- Direct Mexico agronomy lead: “Genotype by environment interaction in sesame germplasm in Mexico” — `10.1002/agj2.70077` (2025), covering rainfed Guerrero environments at metadata level.
- Historical Mexico title — `10.2307/141143` (1941); content relevance requires full-text screening.

No direct regional candidate was found for harvest, postharvest handling, oil extraction, household labor, farm budgets, or food-safety surveillance.

## 5. Food-use safety boundary

Food-safety candidates are included only to define handling boundaries; this dossier makes no medical or therapeutic claim.

- Sesame value-chain food-safety review — `10.1016/j.heliyon.2024.e36347` (2024).
- Sesame seed/oil allergens — `10.1111/j.1398-9995.1996.tb02141.x` (1996).
- Salmonella contamination/inactivation in imported sesame seed — `10.4081/ijfs.2021.8914` (2021).
- Storage structures, mold, and mycotoxin accumulation — `10.3390/toxins15010039` (2023).
- On-farm stored-sesame mycotoxins — `10.3390/agriculture14030372` (2024).
- Fresh/stored sesame mycobiota and aflatoxin B1 — `10.4315/jfp-21-060` (2021).

No home-processing safety thresholds are asserted. Authoritative jurisdiction-specific allergen, microbial, mycotoxin, residue, and sanitation guidance remains required.

## 6. False positives and exclusions

- Sesame extracts/oils applied to bamboo, sunflower oil, orange fruit, cowpea, or meat are not sesame cultivation evidence.
- SESAME/SeSaMe software, solar-cell, access-control, molecular-generation, and scene-editing titles are acronym/homonym noise.
- Medical, clinical, anti-inflammatory, anticancer, diet, and nutraceutical papers are outside crop scope, except narrowly identified food-allergen/safety boundaries.
- Comparative wild *Sesamum* records require explicit species handling.
- Peer-review/decision/response DOI fragments, table/figure supplements, preprint/article pairs, and PDF-suffixed DOI artifacts must not be treated as separate studies.

## 7. Coverage limits and weakest gaps

- The final rerun recorded 30 request failures (10 OpenAlex, 20 Semantic Scholar), while earlier successful raw files remain preserved; failures are explicit in `failures.jsonl` and `search-stats.txt`.
- arXiv and broad keyword searches contributed substantial technical acronym and generic “yield-stress” noise.
- Capsule-development metadata is thin relative to shattering genetics.
- Heat/climate adaptation, soil physical management, organic amendments, integrated disease management, household manual processing, and Mexico economics are weak.
- More than 49% of canonical records lack abstracts, limiting metadata screening.
- rforge v0.1.20 does not expose the documented `evidence grid` or `evidence gaps` subcommands; both attempts returned usage status 2 and are recorded in `evidence-command-errors.txt`. No substitute evidence artifact was fabricated.

## 8. Human gates and next steps

1. **[HUMAN] Screen** `queue.csv` (1,998 pending records), prioritizing the exact candidates above.
2. **[HUMAN] Approve legal/full-text acquisition** for included records.
3. **[AGENT after approval]** create `S-####` notes and extract claims with cultivar, environment, system, product, and loss context.
4. **[HUMAN] Accept extraction, select any analysis method, and approve final scientific claims.**

## 9. Provenance artifacts

- Machine-readable provenance: `provenance.json`.
- Search records: `results.jsonl`, `results-deduped.jsonl`, `results.md`, `raw/`, `manifest.json`, `failures.jsonl`, `search-stats.txt`.
- Resume receipts: `resume-run.txt`, `resume-results.jsonl`, `resume-failures.jsonl`.
- Tool receipt: `tool-version.txt`.
- Coverage: `coverage-stats.txt`, `coverage-analysis.json`, `coverage-analysis.txt`, `candidate-scan.txt`.
- Citation expansion: `citations/*.json`, `citations/*-report.md`, `citation-expansion.txt`.
- Screening queue: `queue.csv`.